Providing a phosphor layer over an LED is common. Typically, a phosphor is provided over a blue LED to make white light. Blue light leaking through the phosphor, combined with the phosphor light, produces white light. There are many ways to provide the phosphor layer over the LED, and one prior art technique is described below.
It is known to pre-form a layer of phosphor powder mixed with silicone and then laminate the layer over blue-emission LED dies mounted on a submount wafer. The wafer is then singulated. The resulting dies emit white light. This is described in United States patent application publications 20110031516 and 20110266569, by Grigoriy Basin et al., assigned to the present assignee and incorporated by reference. Other techniques also mix the phosphor powder in an organic polymer binder (e.g., silicone or epoxy) and then deposit (e.g., print, mold, etc.) the liquid/paste layer directly over the LEDs. The binder is then cured to harden it.
However, the heat and high flux from the LEDs tend to oxidize the organic binder surrounding the phosphor particles, causing the binder to yellow and color shift the light. Further, high quality silicone and epoxy are relatively expensive, which is a significant concern for large remote phosphor components.
What is needed in a process for forming a phosphor layer that can be either formed directly over LEDs or formed on a transparent substrate and which does not use silicone or other organic polymer as a binder.
Although inorganic glass would be a relatively stable and reliable binder for the phosphor powder, molten glass is too chemically reactive at the high temperatures needed to form a phosphor-glass layer, since the glass would react chemically with the phosphors, especially red nitride phosphors.
A sol-gel for forming a glass layer may also be considered as a candidate to substitute for the silicone, but the sol-gel is also too reactive for the phosphor, leading to light attenuation. Further, the low viscosity of sol-gel would result in phosphor sedimentation and non-uniform phosphor density. Other problems would exist as well.